Discussion Paper No.

08-092

Innovation Success
of Non-R&D-Performers
Substituting Technology by Management in SMEs

Christian Rammer, Dirk Czarnitzki,

and Alfred Spielkamp
 Discussion Paper No. 08-092

Innovation Success
of Non-R&D-Performers
Substituting Technology by Management in SMEs

Christian Rammer, Dirk Czarnitzki,

and Alfred Spielkamp

Download this ZEW Discussion Paper from our ftp server:

ftp://ftp.zew.de/pub/zew-docs/dp/dp08092.pdf

Die Discussion Papers dienen einer möglichst schnellen Verbreitung von

neueren Forschungsarbeiten des ZEW. Die Beiträge liegen in alleiniger Verantwortung
der Autoren und stellen nicht notwendigerweise die Meinung des ZEW dar.

Discussion Papers are intended to make results of ZEW research promptly available to other
economists in order to encourage discussion and suggestions for revisions. The authors are solely
responsible for the contents which do not necessarily represent the opinion of the ZEW.
Non-technical Summary

This paper investigates the role of research and development (R&D) and innovation
management for innovation success in small and medium-sized enterprises (SMEs). While
there is little doubt about the significance of in-house R&D activities for generating
successful innovations and obtaining a novelty-based competitive advantage, entering into
continuous in-house R&D activities may be a particular challenge for small firms. They will
have to bear high fix costs such as setting up a separate R&D laboratory, meet minimum scale
requirements of effective R&D activities and have to deal with high sunk costs in case of
stopping R&D. Since the nature of R&D implies high uncertainty of its outcome, devoting
large resources to R&D may jeopardise the whole enterprise in case R&D investment fails.
Given these restraints, SMEs may opt to substitute R&D by innovation managing practices
which demand less investment and bear less uncertainty. Innovation management science has
developed a number of tools which should support firms in identifying innovation potentials,
transferring them into new products and processes and implementing them successfully in the
market or within a firm’s operations. These include human resource management, team
working, co-operating with external partners and sourcing relevant knowledge from the firm’s
environment, particularly from clients, suppliers and competitors.

The paper analyses whether SMEs can in fact substitute R&D activities by innovation
management in order to achieve the same innovation success as R&D performing SMEs. The
research question is closely linked to the innovation policy as it explores whether R&D is a
necessary prerequisite for innovation success, or whether other in-house activities can play a
similar role. The analysis rests on data obtained from the 2003 wave of the German CIS.
Innovation success is measured through a categorial variable that captures the extend to which
an SMEs has successfully introduced “challenging” product and/or process innovations, i.e.
innovations that significantly change the firm’s market position. We run ordered probit
models that control for a potential selection bias between innovating and non-innovating
firms.

Our findings show that continuous R&D activities are a main driver of innovation success in
SMEs, especially when linked to external knowledge sourcing. Firms without in-house R&D
activities can still yield a similar innovation success as R&D performers. On the one hand,
they may contract out R&D which will reduce their own risk and allows them to better control
costs of R&D. On the other hand, human resource management and team work are innovation
management tools that can help non-R&D performing SMEs to gain similar innovation
success as R&D performers, especially when combined with each other or combined with
external knowledge sourcing or formal co-operations with external partners. In contrast,
focusing on searching external sources of innovation without in-house R&D is a less
successful strategy, as is occasional R&D (i.e. starting R&D only in case a technological
problem needs to be solved).
Our results have some relevance for innovation policy. First, the strong focus on promoting
in-house R&D often to be found in innovation policy is not fully supported by our study when
it comes to SMEs. First, in-house R&D seems to be particularly effective only if combined
with external knowledge sourcing. Policy initiatives should thus attempt to combine financial
R&D support to SMEs with strengthening the capacities of SMEs to co-operate with other
partners, including links to customers and suppliers. Secondly, innovation policy should also
acknowledge the key role of external R&D in SMEs and also offer financial support to this
type of R&D activity. Finally, policy may try to identify likely barriers in SMEs that prevent
them from effectively using innovation management practices, particularly human resource
management and team working.
Das Wichtigste in Kürze

In diesem Aufsatz wird die Bedeutung von Forschung und Entwicklung (FuE) sowie des
Innovationsmanagements für den Innovationserfolg von kleinen und mittleren Unternehmen
(KMU) untersucht. Zum einen ist es unstrittig, dass eigene FuE ein wichtiger Erfolgsfaktor im
Innovationsprozess ist, vor allem wenn es um die Entwicklung grundlegender Neuerungen
und um innovationsbasierte Alleinstellungsmerkmale im Markt geht. Zum anderen stehen
gerade KMU vor besonderen Herausforderungen, wenn sie sich kontinuierlich mit FuE
befassen wollen. Hohe Fixkosten etwa im Zusammenhang mit der Einrichtung einer eigenen
FuE-Infrastruktur, Mindestgrößen von FuE-Projekten und hohe sunk costs stellen finanzielle
Barrieren dar. Hinzu kommt, dass FuE von Natur aus mit einem hohen Risiko verbunden ist.
Investieren KMU in beträchtlichem Ausmaß in FuE und bleibt der Erfolg aus, kann dies den
Unternehmensbestand insgesamt gefährden. Angesichts dieser Barrieren verzichten viele
KMU auf eigene FuE-Aktivitäten und versuchen stattdessen, über verschiedene Instrumente
des Innovationsmanagements ihre Innovationsprozesse effizient und effektiv zu gestalten.
Hierzu zählen u.a. personalbezogene Maßnahmen, die Förderung der innerbetrieblichen
Zusammenarbeit zwischen Abteilungen, die Nutzung externer Innovationsquellen und die
Kooperation mit externen Partnern.

Die zentrale Fragestellung dieser Arbeit ist, ob KMU durch ein kluges
Innovationsmanagement eigene FuE-Aktivitäten ersetzen und einen ähnlich hohen
Innovationserfolg wie FuE betreibende KMU erzielen können. Diese Frage ist insbesondere
für die Innovationspolitik relevant, da sie untersucht, ob FuE eine unverzichtbare
Voraussetzung für höheren Innovationserfolg ist - und daher die Förderung an dieser Stelle
ansetzen sollte - oder ob auch andere Strategien Erfolg versprechen. Die empirische Analyse
beruht auf Daten aus der Erhebungswelle 2003 des Mannheimer Innovationspanels.
Innovationserfolg wird durch eine ordinale Variable gemessen, die das Ausmaß abbildet, zu
dem ein KMU „anspruchsvolle“ Innovationen eingeführt hat, d.h. Innovationen, die die
Marktposition des KMU deutlich verbessern. Wir schätzen geordnete Probitmodelle und
kontrollieren dabei für eine mögliche Selektionsverzerrung zwischen innovierenden und nicht
innovierenden KMU.

Die Ergebnisse zeigen, dass eigene FuE tatsächlich ein wesentlicher Erfolgsfaktor für
Innovationen ist, vor allem dann, wenn interne FuE mit der Nutzung von externen
Innovationsquellen (z.B. Kunden- oder Lieferantenimpulse) verbunden wird. Gleichwohl
können Unternehmen ohne eigene FuE unter bestimmten Bedingungen einen ähnlich hohen
Innovationserfolg wie FuE betreibende KMU erzielen. Dies gilt erstens für die Vergabe von
FuE-Aufträgen an Dritte („externe FuE“). Dadurch können die Kosten von FuE gesenkt und
das eigene Risiko verringert werden. Zweitens erweisen sich personalbezogene Maßnahmen
des Innovationsmanagements sowie eine abteilungsübergreifende Zusammenarbeit in
Innovationsprozessen als effektive Instrumente für KMU ohne eigene FuE. Diese Instrumente
sind insbesondere dann erfolgreich, wenn sie miteinander oder mit der Nutzung externer
Innovationsquellen bzw. Kooperationen mit externen Partnern kombiniert werden.
Demgegenüber erzielen KMU ohne eigene FuE, die vorrangig auf die Nutzung externer
Innovationsquellen als Innovationsmanagementstrategie setzen, einen deutlich niedrigeren
Innovationserfolg als forschende KMU. Ebenfalls niedrigere Innovationserfolge erzielen jene
KMU, die nur gelegentlich (d.h. anlassbezogen) forschen.

Aus unseren Ergebnissen kann für die Innovationspolitik der Schluss gezogen werden, dass
Maßnahmen, die alleine auf die Förderung von interner FuE setzen, nicht notwendigerweise
der beste Weg sind, um die Innovationserträge der Unternehmen und damit auch die
gesamtwirtschaftlichen positiven Effekte von Innovationsaktivitäten zu maximieren. Für
KMU gilt, dass interne FuE mit der Nutzung von externen Innovations- und
Technologiequellen verknüpft werden sollte. FuE-Förderung an KMU sollte daher mit
Kooperationsansätzen, vor allem auch was die Einbeziehung von Kunden und Lieferanten
betrifft, verbunden werden. Des Weiteren könnte der positive Effekt der Vergabe von FuE-
Aufträgen dadurch gestärkt werden, dass diese Form der FuE-Tätigkeit mit interner FuE in
der staatlichen Förderung gleichgestellt wird. Schließlich sollte festegestellt werden, welche
Barrieren KMU davon abhalten, ein effektives Innovationsmanagement einzuführen,
insbesondere im Hinblick auf personalbezogene Maßnahmen und abteilungsübergreifende
Zusammenarbeit.
 Innovation Success of Non-R&D-Performers
Substituting Technology by Management in SMEs

Christian Rammera, Dirk Czarnitzkib,a and Alfred Spielkampc,a

October 2008

Abstract

This paper investigates the impact of in-house R&D and innovation management practices
on innovation success in small and medium-sized firms (SMEs). While there is little doubt
about the significance of technology competence for generating successful innovations, in-
house R&D activities may be a particular challenge for SMEs due to high risk exposure, high
fixed costs, high minimum investment and severe financial constraints. SMEs may thus opt
for refraining from R&D and relying more on innovation management tools in order to
achieve innovation success. We analyse whether such a strategy can pay off. Based on data
from the German CIS we find that R&D activities are a main driver for innovation success if
combined with external R&D, using external innovation sources or by entering into co-
operation agreements. SMEs without in-house R&D can yield a similar innovation success if
they effectively apply human resource management tools or team work to facilitate innovation
processes.

Keywords: Innovation Success, R&D, Innovation Management, SMEs

JEL-Classification: L25, O31, O32, O38, O47

Acknowledgements: The authors would like to thank Marco Vivarelli, Roy Thurik,
Jacques Mairesse and the other participants of the workshop
„Drivers and Impacts of Corporate R&D in SMEs“, held at the JRC-
IPTS in Seville, September 19th, 2008 as well as Bettina Peters for
helpful comments. The usual disclaimer applies.

a
Centre for European Economic Research (ZEW), Department of Industrial Economics and International
Management, L 7, 1 - D-68161 Mannheim; rammer@zew.de (corresponding author).
b
K.U. Leuven, Dept. of Managerial Economics, Strategy and Innovation, Naamsestraat 69, B-3000 Leuven;
dirk.czarnitzki@econ.kuleuven.be.
c
University of Applied Sciences Gelsenkirchen, Department of Entrepreneurship and Innovation,
Neidenburger Str. 43, D-45877 Gelsenkirchen; alfred.spielkamp@fh-gelsenkirchen.de.
1 Introduction
The ability to generate new knowledge by research and development (R&D) is generally
regarded as a key driver for innovation success of firms. R&D findings are likely to result in
superior product characteristics or a significant increase in production efficiency which can be
transferred into higher market success of new products or new processes. Empirical research
could show a positive impact of in-house R&D on innovation success in terms of sales of new
products or the degree of novelty (see Cassiman and Veugelers, 2005; Becker and Peters,
2000). High investment in R&D by SMEs was also found to increase patent outcome and total
factor productivity (see Cohen et al., 1987; Rogers, 2004; Lee and Sung, 2005; Plehn-
Dujowich, 2007; Kim et al., 2004; Dijk et al., 1997). At the same time, R&D is a costly and
risky activity that demands a minimum amount of resources and time in order to achieve
results. SMEs may face particular barriers while investing in in-house R&D which add to the
general constraint for private R&D investment resulting from knowledge spillovers. High
fixed costs and a high minimum size of R&D projects demand a high financial commitment,
put pressure on profits and imply high risk on firm survival in case an R&D project fails to
deliver. Financing R&D is particularly challenging for SMEs as loans are rarely available due
to high risk exposure (i.e. not bankable risk) and lack of collaterals while both loan and equity
financing are hampered by information asymmetries as well as high management costs on the
side of the financer in order to evaluate and monitor investment into SMEs (see Hall, 2002).
Given these constraints, SMEs may refrain from continuous R&D activities. In order to still
benefit from the opportunities of innovation, they could try to invest in less risky and costly
activities to generate innovations, particularly by putting more emphasis on managing
innovation processes and fully exploiting their innovative potentials by appropriate
management techniques, including the use of external knowledge.
The aim of this paper is to empirically analyse whether such a strategy pays off. For this
purpose we distinguish between technological competence, organisational skills for managing
innovation and network competence as driving forces for innovation success (see Ritter and
Gemünden, 2004). While technological competence is closely related to in-house R&D,
organisational skills refer to practices for smoothly organising innovation processes, including
intra-firm communication and incentives for employees to actively contribute to innovation
efforts (see Tidd et al., 2005; Hidalgo and Albors, 2008). Network competence relates to a
firm’s ability to use external sources of innovation and incorporate them into the internal

6
innovation processes. In practice, a firm will have to combine all three fields of competence
to successfully innovate. In this paper, we are particularly interested in whether SMEs that
decide not to engage in in-house R&D can compensate for the resulting deficiency in
technological competence by investing more into organisational skills and network
competence. From this reasoning, three research questions emerge:
(1) Are SMEs that invest into in-house R&D more successful in their innovative efforts
than innovating SMEs without in-house R&D?
(2) Are there innovation management strategies that allow non-R&D performing SMEs to
achieve similar innovation success compared to SMEs with in-house R&D?
(3) What elements do successful innovation management consist of?
These questions are relevant to innovation policy targeting small businesses. Since policy
attempts to support successful innovation as a means to increase wealth, it is pertinent to
know what barriers prevent firms from successful innovation, and what type of public support
is needed to maximise innovation success. Innovation policy in most industrialised countries
assumes that supporting R&D is critical in this respect, both in small and large firms. This
policy approach may be challenged in case the link between R&D and innovation success is
weak in SMEs, and innovation success is actually driven by sophisticated innovation
management.
In the next section, we discuss the role of R&D and innovation management in SMEs with a
particular focus on the likely constraints SMEs may face when engaging in R&D. Section 3
discusses our approach to measuring innovation success in SMEs and section 4 presents our
empirical model. The data used are described in section 5 while section 6 summarises the
model estimation results and discusses main findings. Section 7 concludes this paper.

2 R&D and Innovation Management in SMEs

R&D is by no doubt a major ingredient to innovation. Exploring new ways of solving
technical problems, employing new technologies to meet user demand and developing new
technologies to produce and deliver goods and services will help firms generate innovations
that outperform competitors and help innovators to gain market shares and increase
profitability. Yet, investing into R&D is associated with high costs and risk. Any firm will
thus have to balance between the expected benefits from successful R&D and the costs and
probability of failure when engaging in R&D. There are several features of R&D that are

7
likely to result in systematic differences between small and large firms with respect to
conducting in-house R&D:
- R&D is subject to minimum project sizes in order to generate useful results due to
technical indivisibilities. SMEs will thus have to invest a higher share of their resources
to R&D than large firms, limiting their ability to invest in other business areas such as
marketing. Minimum costs may even be as high to prevent SMEs from any R&D
investment (see Galbraith, 1952: 92).
- R&D is associated with high entry costs, i.e. specific investment into laboratory
equipment and human capital. Given that R&D is subject to technical indivisibilities,
SMEs will have to invest a high share of their total sales to set up R&D. In case of
stopping R&D, this investment is likely to be sunk costs.
- R&D costs are largely fixed costs. SMEs have to spread these fixed costs over a smaller
sales base than large firms, militating either profitability or price competitiveness and
restricting available cash flow to finance R&D in the future (see Cohen and Klepper,
1996).
- R&D activities are highly idiosyncratic. Outsiders such as external financers will find it
difficult to evaluate the prospects of a firm’s R&D efforts in terms or risk and technical
and economic potential without a firm-specific history of success and failure over a
larger number of projects. SMEs, and especially young firms, are less able to provide
such a track record and may suffer restricted access to external finance.
- Most R&D is investment, i.e. returns, if any, are generated in later periods than
expenditures occur. While R&D demands pre-financing, most R&D is current
expenditures for staff and material and does not qualify as collaterals for debt financing.
Lack of collaterals and information asymmetries on R&D project perspectives raise
costs of external funding (see Freel, 2007; Czarnitzki, 2006; Tiwari et al., 2007).
Financing in-house R&D thus strongly rests on internal funding sources. Their
availability tends to be more restricted in SMEs due to a smaller cash flow.
- R&D is risky, and many R&D projects fail. While large firms are able to spread risk by
running a portfolio of different R&D projects at the same time, SMEs will have to focus
on one or a few projects only. Failure of a single R&D project may increase the risk
exposure of the firm as a whole substantially. This is particularly relevant when SMEs
finance R&D through debt using non-R&D assets as collateral. Liquidating these assets
in case of failure of R&D projects may jeopardise the entire business.

8
 These features of R&D typically result in a lower propensity of SMEs to conduct R&D.
This is regularly revealed by data from innovation surveys which cover a random sample of
firms from various size classes (see Kleinknecht, 1989; Santarelli and Sterlacchini, 1990).
Current data from the German Innovation Survey conducted in 2007 (see table 1) show that
across a large set of manufacturing and service sectors, the share of firms with in-house R&D
activity on a permanent basis increases steadily by size class, from 8.2% (firms with 5 to 9
employees) to 52.3% (500 and more employees). The same holds true when the sector
coverage is restricted to sectors where competition is particularly based on research,
innovation and new knowledge. “Permanent R&D” refers to R&D activities that are
performed continuously and independently from actual demand for developing or improving
technology. Permanent R&D activities typically imply a certain number of staff which is
solely assigned to R&D, and investment into dedicated laboratory equipment.

Table 1: Share of firms engaging in in-house R&D and in any type of innovation activities,
by size class (2004-2006, per cent of all firms)
All sectors1) R&D and knowledge intensive sectors2)
Size class in-house in-house other any type in-house in-house other any type
(# employees) R&D, on R&D, types of of inno- R&D, on R&D, types of of inno-
a perma- only inno- vation a perma- only inno- vation
nent base occasio- vation activity nent base occasio- vation activity
nally activities (incl. nally activities (incl.
R&D) R&D)
5 to 9 8.2 10.2 27.0 45.4 13.8 10.5 30.7 55.0
10 to 19 9.3 10.2 35.0 54.5 17.5 12.2 35.2 64.9
20 to 49 13.2 14.4 28.2 55.8 26.2 15.6 26.5 68.4
50 to 99 15.1 19.5 30.7 65.2 30.5 22.0 28.0 80.5
100 to 249 24.2 16.1 30.7 70.9 47.4 14.3 24.2 86.0
250 to 449 37.4 10.6 24.5 72.5 58.6 12.7 18.8 90.0
500 and more 52.3 7.8 25.6 85.6 74.8 5.5 15.4 95.7
Total 12.0 12.0 29.5 53.5 20.5 12.5 30.4 63.3
Note: R&D and other types of innovation activities refer to the three year reference period 2004-2006.
1) NACE (rev. 1.2) 10-41, 51, 60-67, 72-74, 90, 92.1-92.2 - 2) NACE (rev. 1.2) 24, 29-35, 64.3, 65-67, 72-73, 74.1-74.4, 92.1-92.2
Source: ZEW - German Innovation Survey 2007; weighted figures.

While most SMEs refrain from conducting R&D permanently, a large share of SMEs still
conducts innovation activities since this may help them to compensate for various other
liabilities of smallness (see Brüderl and Preisendörfer, 1998). About a third of innovative
firms with no permanent in-house R&D perform R&D on an occasional basis. They devote
resources to R&D only in case there is a direct demand from other business functions such as
production or marketing. This strategy reduces fixed costs of R&D and limits funding
requirements. A main drawback may be less sophisticated R&D outputs in terms of
technological advancement and novelty since occasional R&D restricts the resources to

9
continuously monitor relevant technology trends and reduces the capacity to absorb external
knowledge.
The majority of innovative firms without permanent in-house R&D refrain from any kind of
R&D activity. These firms may rest on technology and knowledge inputs from external
sources such as suppliers or consultants. But they may also have found ways other than R&D
to exploit internal innovation potential and access external sources for innovation. In
particular, sophisticated innovation management methods could help firms to achieve similar
results with innovations and market success as R&D performing firms do. Innovation
management covers a broad set of tools and techniques (see Nijssen and Frambach, 2000;
Hidalgo and Albors, 2008). We use the term here to denote all activities of firms targeted at
organising the innovation process in a way to maximise the outcome in terms of market
success with new products and new processes. Innovation management includes measures to
facilitate both internal processes and external links. Facilitating internal processes include
organisational skills for identifying innovation ideas, providing incentives to managers and
improving co-operation among business units and departments. Monitoring R&D and
innovation projects with a particular focus on identifying a project’s prospects (and stopping
those that will not deliver) is another key element of innovation management.
Improving external links requires network competence such as techniques to identifying
innovation impulses from customers, leveraging suppliers as source of innovation or
absorbing knowledge from other organisations, including competitors and public research.
Entering into R&D co-operations, research joint ventures or other forms of partnerships are
popular ways to access knowledge available at other organisations. Identifying, accessing and
absorbing external knowledge requires certain capabilities often associated with conducting
in-house R&D (Cohen and Levinthal, 1989, 1990; Rosenberg, 1990). Balancing between
internal R&D efforts and external knowledge acquisition is thus another major concern of
innovation management (see Arora and Gamardella, 1994; Cassiman and Veugelers, 2005). In
recent years, the notion of “open innovation” (Chesbrough, 2003; Roper et al., 2007) has
directed attention to the management of various external links in innovation, including human
capital and finance.
While innovation management activities of firms may be categorised in many different
ways (see Tidd et al., 2005; Adams et al., 2008), we distinguish four areas of innovation
management practice through which firms may improve their innovative performance, the
first two are related to organisational skills, the latter two to network competence:

10
 - Human resource management (HRM) aims at increasing incentives for managers and
employees to engage in innovation activities and develop skills needed for effective
innovation efforts. In particular, HRM needs to motivate and enable individuals to
experiment with new ideas (Shipton et al., 2005). Management practices include
recruiting methods to identify the right people for promoting innovation within an
organisation, training for handling innovation challenges, as well as reward systems,
performance management systems and career development tools that help in the
formation of innovative ideas of employees.
- Team working is intended to facilitate knowledge sharing, develop mutual trust and help
overcoming organisational barriers. The role of teams for increasing performance has
first been shown for production (Levine, 1995) and quality management (see Hoegl and
Gemuenden, 2001). With regard to innovation, creating cross-functional teams has been
identified as particularly important for accelerating innovation processes from R&D to
marketing (see Zeller, 2002) and smoothing information flows among different business
units (Allen, 1983). The importance of cross-functional co-operation for innovation
success was shown by Cooper and Kleinschmidt (1995), Love et al. (2006) and Song et
al. (1997). Management tools range from joint workshops, knowledge information
systems for open cross-functional communication and innovation circles to initiate a
temporary exchange of personnel across units.
- Searching external sources of innovation aims at identifying valuable impulses from
customers, suppliers, competitors or universities and other public research organisations
in order to orient innovation efforts (see Katila and Ahuja, 2002; Laursen and Salter,
2006). Key management issues refer to methods of identifing key innovation sources
such as lead customers or lead markets (see von Hippel, 1998; Beise, 2003), to assess
the value of external sources and to develop in-house capacities to incorporate these
impulses into the innovation processes.
- Co-operation in innovation and other forms of partnering and external knowledge
acquisition should provide access to complementary knowledge (see Baumol, 2002) and
can help sharing the costs and the risk of innovative activities (Hagedoorn, 2002). Co-
operations typically rest on formal agreements. Managing intellectual property issues
and the distribution of costs and returns of joint innovative efforts are key management
issues of this type of innovation practice.
Results from the German Innovation Survey in 2003, which contained a large set of
questions on innovation management practices, show that each of the four above mentioned

11
areas is widely used (see table 2).4 72% of innovative firms in Germany effectively use team
work, i.e. these practices were assessed as highly important for facilitating innovation
processes. A similar share effectively employs searching for external innovation sources, i.e.
they were able to identify these sources and transfer external impulses into new products or
processes. HRM is effectively used by 57% of all innovative firms, and 48% engage in co-
operation agreements for developing innovations.

Table 2: Share of innovative firms effectively using innovation management practices, by

size class (2000-2002, percent of all innovative firms)
All sectors1) R&D and knowledge intensive sectors2)
Size class HRM Team Search Co-ope- HRM Team Search Co-ope-
(# employees) work ration work ration
5 to 9 53 69 75 48 56 65 76 40
10 to 19 50 65 58 41 52 71 63 40
20 to 49 67 77 72 53 60 78 69 56
50 to 99 59 81 70 50 63 84 71 37
100 to 249 62 81 67 45 69 82 70 47
250 to 449 71 85 77 53 73 84 77 58
500 and more 80 86 82 71 85 90 83 76
Total 57 72 71 48 58 72 72 44
Note: Effective use of innovation management practices refers to the three year reference period 2000-2002.
1) NACE (rev. 1.2) 10-41, 51, 60-67, 72-74, 90, 92.1-92.2 - 2) NACE (rev. 1.2) 24, 29-35, 64.3, 65-67, 72-73, 74.1-74.4, 92.1-92.2
Source: ZEW - German Innovation Survey 2003; weighted figures.

While large innovative firms show the highest shares for all four areas, size differences in
the effective use of innovation management practices within the group of SMEs are marginal.
This suggests that there are little if any size-related barriers to apply innovation management
techniques successfully. What is more, innovation management practices are widespread in
all sectors. Firms in research and knowledge intensive sectors do not show a significantly
stronger use of any of the four practices. This pattern is in stark contrast to the finding for
permanent in-house R&D.

3 Innovation Success in SMEs

The key purpose of our analysis is to compare the level of innovation success among
innovative SMEs with varying in-house R&D activities and innovation management
practices. A proper measure of innovation success has to respond to three challenges. First,
the measure must represent both product and process innovation success, since R&D and

4
Definition and measurement of the four areas of management practice is described in section 4.

12
innovation management activities can target both types of innovations. Secondly, the measure
should distinguish between more and less successful innovators. Thirdly, the measure has to
be neutral to firm size.
Measures of innovative output such as patents or sales with new products which are
typically used in empirical studies on innovation success (see Kleinknecht et al., 2002) do not
fully meet these requirements. While patents are clearly a valid measure for the success of
R&D efforts (see Griliches, 1984; Hall et al., 1986, 2001), output of other types of innovative
activity may be incompletely captured by this indicator. What is more, smaller firms may be
reluctant to use patents due to high patenting costs and a lack of experience in successfully
defending their intellectual property rights against offense (see Soete, 1979; Acs and
Audretsch, 1988, 1991; Licht and Zoz, 2000). Another drawback of patents as an indicator of
success is the lack of market evaluation since patents give no indication whether the new
technological knowledge protected by patents was transferred into market success.
Studies based on innovation survey data most often use the share of sales generated by new
products as a success indicator (see Crépon et al., 1998; Belderbos et al., 2004; Lööf and
Heshmati, 2002; Janz et al., 2004; Griffith et al., 2006; Parisi et al., 2005; Roper et al., 2007),
sometimes distinguishing between new-to-the-market and new-to-the-firm products (see de
Jong and Vermeulen, 2006). The main drawback of this indicator is its sole focus on product
innovation success. Some authors argue that most process innovation is related in some way
to product innovation (Mohnen and Röller, 2005), but there are clearly many successful
process innovations that are not related to product innovation success, particularly with regard
to cost saving process innovation.
In this study, we propose an alternative measure of the level of innovation success in
innovative SMEs that meets all three requirements. We focus on an SME’s ability to
introduce “challenging” innovations, i.e. innovations that demand particular efforts and are
likely to alter an SME’s market position very significantly. We furthermore suppose that a
firm which is able to introduce more “challenging” innovations at the same time is more
successful than one introducing less.
For product innovation, we propose that innovation success in terms of an innovation
effects on a firm’s market position basically depends on the degree of novelty. A high degree
of novelty will result in a more pronounced product differentiation vis-à-vis competitors, is
more likely to gain new customers and will allow a firm to realise higher mark-ups. A high
degree of novelty is especially relevant for SMEs in order to distinguish themselves from
larger competitors and compensate for liabilities of smallness such as a lack of reputation and

13
less resources for marketing new products. There are two types of novelty to be distinguished:
A market novelty is a new product that has not been supplied in the same or a similar form on
a firm’s market yet. Market novelties need not be world’s first necessarily since a firm’s
market may be restricted to a certain region or group of customers. Another type of novelty
refers to a firm’s range of products. “Product line novelties” are new products that have no
predecessor product within the firm. They allow firms to enter a new market or market
segment. While a product line novelty may either be a market novelty or an imitation of
products offered by other firms in the respective market (segment), it challenges a firm’s
management as it may demand new ways of production, distribution and marketing.
Introducing both market and product line novelties at the same time means that SMEs will
have to cope with new market environments and will have to convince their customers of new
product features their are not familiar with yet. Yet, a new product may be a market novelty
and a product line novelty at the same time in case a firm enters a new market by offering a
product that has not been offered at that market by any other firm before. In this case, a firm
is taking particularly high risk which may be compensated by particularly high gains in
market shares and profits. If having introduced such an innovation successfully it is fair to
consider this firm as being more successful with introducing challenging innovation than a
firm with a market novelty or a product line novelty only.
For process innovation, we argue that a successful introduction of new processes can
basically yield to two types of outputs: a decrease in unit costs of production, and an increase
in the quality of production processes. A new process that reduces unit costs may be called
“efficiency innovation” and is likely to increase price competitiveness of firms, resulting in
higher profits or higher market shares (see Peters, 2008). Quality improving process
innovation (“quality innovation”) increases may allow for product differentiation which could
have a positive impact on a firm’s competitiveness. While efficiency innovations often focus
on automation, simplifying procedures and realising synergies, quality innovations are likely
to demand more precise and sometimes more time consuming and costly production
processes. Quality innovations constitute a separate output dimension since they aim at
increasing product and service quality (of both new and old products), resulting in higher
sales either through an increase in demand or through higher product prices. Quality
innovations are particularly relevant in services since many service innovations rest on
reshaping processes in a way to better respond to new customer needs (see Miles, 2005).
Combining both outcomes of process innovation indicates a particularly challenging type of
process innovation activity.

14
 All four types of innovation success may be measured through qualitative and quantitative
indicators. Qualitative indicators simply indicate whether a firm has introduced a certain type
of innovation during a given period of time. Quantitative indicators capture the significance of
these innovations in a firm’s total activities. We use the sale share of market novelties and
product line novelties as well as the share of unit costs reduced by efficiency innovations and
the increase in sales due to quality innovations as indicators of quantitative innovation
success. In order to obtain a single success measure, we construct a simple index following
Bresnahan et al. (2002) by applying a z-transformation to each indicator (i.e. an indicator’s
mean over the whole sample is subtract from a firm’s value of this indicator, divided by the
standard deviation) and summing up the transformed values.
A main drawback of this indicator is its size dependence. When looking at the group of
firms that have successfully introduced a certain type of innovation, quantitative innovation
success is systematically higher for very small firms compared to medium-sized or large firms
(see table 4). One may argue that very small firms are particularly successful innovators and
much more capable to generate high sales shares, cost reductions and sales growth from
innovations than large firms. We rather believe that the result shown in table 3 reflects effects
of small numbers which makes it more likely for very small firms to obtain high indicator
values due to small base values (i.e. a low volume of sales of a low amount of total costs).

Table 3: Innovation success indicators for innovating firms (2002, per cent)
Size class Sales share of Sales share of Unit cost reduction Sales increase due
(# employees) market novelties1) product line through efficiency to quality
novelties1) innovations1) innovations1)
mean std. dev. mean std. dev. mean std. dev. mean std. dev.
5 to 9 27.8 26.2 28.2 27.2 18.3 15.3 16.6 12.5
10 to 19 26.2 27.8 24.9 26.4 10.6 6.6 18.8 20.5
20 to 49 22.0 26.1 24.0 27.6 10.9 10.8 14.9 16.6
50 to 99 18.5 22.3 16.7 19.9 9.3 7.3 12.9 13.9
100 to 249 14.3 18.1 14.5 20.5 8.4 6.6 10.0 12.0
250 to 449 11.9 12.7 9.4 8.2 7.2 5.6 6.5 5.2
500 and more 13.0 18.5 10.0 14.7 8.2 9.1 6.5 7.8
Total 18.8 22.9 18.3 23.0 9.6 9.2 12.1 14.1
Note: Firms from NACE (rev. 1.2) 10-52, 60-74, 90, 92.1-92.2.
1) Only firms which had introduced the respective type of innovation during 2000-2002.
Source: ZEW - German Innovation Survey 2003; sample means.

One may response to this problem by either calculating size class specific success indicators
or by controlling for size class effects in econometric modelling. There is still another
drawback, however, which refers to effects of a firm’s number of different products and
processes on the level of innovation success as measured by the above mentioned indicators.

15
Suppose an SME that just offers one product which is replaced by a new product which is
new to the firm’s market. If this new product is successfully introduced, it will generate a
sales share of 100 per cent. In contrast, a firm which offers two products (each generating half
to total sales) and which replaces one of its products by a market novelty successfully will
report a sales share with market novelties of 50 per cent. One may argue, however, that both
have obtained a similar degree of innovation success since both were able to develop a market
novelty and introducing it to the market. A similar argument can be made with respect to sales
growth due to quality innovations and the magnitude of cost savings from efficiency
innovations which are also likely to be higher the lower the number of different processes a
firm applies.
Since our data do not contain information on a firm’s number of different products and
processes we are not able to control for a resulting bias in the innovation success indicator.
We thus refrain from applying a quantitative indicator but instead use a count variable which
counts the number of different innovations (market novelties, product line novelties,
efficiency innovation, quality innovation) introduced within a certain period of time. We
argue that an SME that was able to introduce both market and product line novelties and at
the same time implement new process technologies that yield to both cost savings and quality
improvements is obviously successful in reshaping its market position substantially through
innovation, both in terms of product differentiation, entering new markets and improving
price competitiveness. While the immediate quantitative impact of these innovations in terms
of sales shares, cost savings or sales growth may vary considerable among individual firms,
these differences do not adequately reflect the likely long term effects of such innovations. On
the other hand, firms having introduced innovations that were neither market novelties nor
product line novelties, efficiency innovations or quality innovations will obtain significantly
smaller returns from their innovative activity. These firms did introduce product imitations
(i.e. new products in their established market which were offered by competitors already
before) or process innovations that did not yield any significant changes in costs or quality.
For firms with one to three different types of “challenging” innovation, we assume that
innovation success always increases when another type of innovation is successfully
introduced, regardless of the type. While one may argue that some of the four types
innovations are more valuable than others, particularly with regard to market novelties (which
may give a firm temporary monopoly), we stress the strategic importance for an SME being
able to compete on several dimensions of innovations. It clearly increases a firm’s ability to
response to changing market environments and competitive strategies of competitors, and it

16
reduces an SME’s dependence from one particular innovation. We thus treat our innovation
success indicator as a categorical variable which reflects the degree of complex innovation an
SME is capable to introduce at the same time.
Results from the most recent innovation survey in Germany show that out of all firms with
product or process innovations within a three year period, 29% introduced market novelties,
30% product line novelties, 38% efficiency innovations and 41% quality innovations (see
table 4). While large firms show the highest shares throughout, size effects seem to be more
prevalent for product line novelties and less for efficiency and quality innovations. Only 5%
of all innovators were able to succeed with all four types. 10% could introduce three different
types within a three year period. About a quarter of all innovators introduced none of the four
types, i.e. they either introduced mere imitations or new processes that did not result in lower
unit costs or higher product quality.

Table 4: Share of innovative firms successfully introducing different types of innovations

(2004-2006, per cent)
Type of innovation1) Number of different types of innovation2)
Size class Market Product Efficien- Quality 0 1 2 3 4
(# employees) novelty line cy inno- inno-
novelty vation vation
5 to 9 23 19 36 27 35 34 22 8 1
10 to 19 27 33 28 49 27 25 36 10 2
20 to 49 32 31 40 47 21 31 30 11 7
50 to 99 37 32 47 44 20 27 34 11 8
100 to 249 37 40 48 44 15 35 26 15 9
250 to 449 39 52 47 56 12 21 38 17 12
500 and more 52 63 62 59 9 20 24 20 27
Total 29 30 38 41 26 30 29 10 5
Note: Firms from NACE (rev. 1.2) 10-41, 51, 60-67, 72-74, 90, 92.1-92.2.
1) Share in all firms with product or process innovations in 2004-2006.
2) Share in all firms with product or process innovations in 2004-2006; “0” indicates that firms have introduced new products or processes
that neither were market novelties nor product line novelties, efficiency innovations or quality innovations. “4” indicates that firms have
introduced all four types.
Source: ZEW - German Innovation Survey 2007; weighted figures.

4 Empirical Model
For investigating our research questions, we analyse the effects of conducting R&D and
using various types of innovation management tools on an SMEs’ success in introducing
complex innovations. We start from a base-line model that links innovation success SUC (i.e.
the index described above) to R&D input, innovation management efforts and a set of k
control variables (CTR). R&D input in SMEs can relate to in-house R&D and external R&D
(i.e. contracting out R&D to other firms or universities). Following Cassiman and Veugelers

17
(2006), we allow that combining in-house and external R&D may have different effects than
doing only in-house or only external R&D. A firm’s i R&D input is thus measured by three
dummy variables: in-house R&D conducted on a permanent base (RI), external R&D (RE)
and an interaction term (RI*RE).5
Since SUC can be observed for innovating firms only, one has to take into account a
potential selection bias. When analyzing SUC we perform regressions conditional on the fact
that the firm is an innovator. If innovating and SUC are correlated, the estimates using only
innovators would be biased. Hence, we apply a sample selection model where we first model
the decision to be innovative and second analyze SUC. As SUC is an ordinal variable we
cannot apply the well-known sample selection model by Heckman (1979). Due to the non-
linearity of the Ordered Probit model, we have to estimate the two equations jointly by Full
Information Maximum Likelihood (FIML) where we allow error term correlations across the
two equations. Thus, a selection equation estimates the probability that firm i introduced an
innovation INN depending on a set of explanatory variables (EXV) while a success equation
models the effects of R&D input and a set of control variables CTR on the level of innovation
success of an innovating firm (SUC). The baseline model thus reads:

INNi* = ζ + Σj δj EXVij + νi , [1a]

SUCi* = α + β1 RIi + β2 REi + β3 RIi*REi + Σk δk CTRik + εi , [1b]

where we observe INN = 1 if INN* > 0 and INN = 0 otherwise. SUC is only observed if
INN = 1, and takes values from 0 to 4. Assuming a joint normal distribution of the two error
terms allows us to estimate the model by familiar FIML techniques. For technical details, see
Miranda and Rabe-Hesketh (2006).
Following the literature on a firm’s propensity to innovate (see Crepon et al., 1998; Cohen,
1995; Acs and Audretsch, 1988; Bhattacharya and Block, 2004), we use firm size, firm age, a
firm’s ability to absorb knowledge (measured by the share of skilled labour), the market
environment (captured by the significance of export markets for a firm’s total sales) and

5
In section 2, we argued that conducting in-house R&D occasionally is another type of R&D input which may
be preferred by SMEs since it involves less investment, less financial resources and bears less risk. As this paper
is about whether SMEs need to invest into R&D as a permanent activity in order to achieve high innovation
success or whether they can substitute R&D by some type of innovation management, we refrain from
considering occasional R&D in the remainder of the paper. Estimation results of extended models that also
included a term for occasional R&D and an interaction term for occasional and external R&D showed that
neither of the both terms was statistically significant, indicating that occasional R&D does not support
innovation success in SMEs in our empirical set-up.

18
whether a firm belongs to an enterprise group as explanatory variables EXV in [1a]. The
control variables CTR in [1b] include size and age (the latter controlling for likely effects of
market and technology experience), the financial input devoted to innovative activities (as a
share in total sales), receipt of public subsidies for innovation (which is likely to increase the
available funds for innovation at almost zero costs and should thus enable the firm to devote
larger efforts to developing and successfully introducing challenging innovations) and lagged
export intensity (as a proxy for the competitive environment, assuming stronger competitive
pressure on firms substantially engaged in foreign markets). Furthermore, we control for
sector affiliation and region.
In a further step, we add a firm’s innovation management practice to [1b], the selection
equation [1a] remaining unchanged. Dummy variables capture whether a firm effectively uses
human resource management (HRM), team work and cross-functional co-operation within the
firm (TMW), searching for external sources of innovation (SEA), and co-operation with
external partners to develop innovations (COP).

SUCi = α + β1 RIi + β2 REi + β3 RIi*REi + χ1 HRMi + χ2 TMWi

+ χ3 SEAi + χ4 COPi + Σk δk CTRik + εi [2b]

We assume that engaging in in-house R&D activity is a basic decision any innovative SME
will have to make. Since SMEs are faced with specific constraints to enter into this activity,
some SMEs will opt to not perform R&D. In order to identify whether these firms can achieve
similar innovation success through focusing on innovation management, we interact RI and
the four variables of innovation management, thus distinguishing between firms conducting
both in-house R&D and applying certain innovation management practices at the same time,
and those using innovation management without in-house R&D:

SUCi = α + β1 RIi + β2 REi + β3 RIi*REi + χR1 HRMi*RIi + χR2 TMWi*RIi +

χR3 SEAi*RIi + χR4 COPi*RIi + χN1 HRMi*(1-RIi) + χN2 TMWi*(1-RIi) +
χN3 SEAi*(1-RIi) + χN4 COPi*(1-RIi) + Σk δk CTRik + εi [3b]

By keeping RI as a separate variable in the model, we can test whether conducting in-house
R&D adds to innovation success when controlled for interaction effects between R&D and
innovation management. A positive and statistically significant value for β1 in [3b] suggests
that R&D performing firms have an advantage in generating complex innovations over non-
R&D performing firms which cannot be fully compensated through innovation management.

19
[3b] can also be used to apply a simple test of substitution effects between in-house R&D and
the four innovation management practices on innovation success. Substitution is present if

(β1 + χRm) ≤ χNm for m = {1,...,4}, [4]

that is, the combined effect of conducting in-house R&D jointly with a certain innovation
management practice m is statistically not significantly higher than the effect of applying
innovation management practice m without in-house R&D.
Aside from a potential substitution of R&D by innovation management practices, we are
also interested in exploring the most successful ways of managing innovation processes in
SMEs. For this purpose, we construct dummy variable for different “innovation management
practices” (IMP) for all 16 combinations n of HRM, TMW, SEA and COP. Innovation
management practice is multiplied by RI and (1 - RI) in order to separate practices in R&D
and non-R&D performing firms.

SUCi = α + β1 RIi + β2 REi + β3 RIi*REi + Σn χRn IMPin*RIin +

Σn χNn IMPin*(1-RIin) + Σk δk CTRik + εi [5b]

For each combination of innovation management practices, we can then test whether
internal R&D plus a certain management practice combination leads to a higher output
compared to the situation where a firm applies this practice without conducting permanent
R&D. Furthermore, the test for substitution effects [4] can be applied accordingly.

5 Data
The study rests on data from the German Innovation Survey, which is the German
contribution to the EU’s Community Innovation Survey (CIS). While the German Innovation
Survey fully complies with the methodological recommendations for CIS surveys and adopts
the standard CIS questions, it goes beyond the CIS design in three important respects (see
Janz et al., 2001, for a more detailed discussion). First, the German Innovation Survey is
designed as a panel survey and is conducted every year. Every year the same gross sample of
firms is surveyed, refreshed biannually to compensate for panel mortality. The Survey is
conducted by the Centre for European Economic Research (ZEW) located in Mannheim, thus
also known as Mannheim Innovation Panel (MIP). Secondly, the MIP contains a significantly
larger number of questions compared to the harmonised CIS questionnaire, which allows for a
much more in depth analysis of relations between firms’ innovation activities, their market

20
environment and their economic performance. Thirdly, the MIP has a somewhat broader
sector and size coverage than the CIS standard, including firms with 5 to 9 employees and
covering a larger set of service sectors.
This paper uses data from the survey wave 2003. In this year, the questionnaire contained a
number of special questions related, among others, to the use of external innovation sources
and innovation management practices in the areas of human resource management and cross-
functional teams. The gross sample of the survey was 25,791 firms, from which 3,272 were
classified as neutral losses due to firm closure, mergers and acquisitions or other events. 4,583
firms responded to the survey, which equals 20.2% of the gross sample corrected for neutral
losses. The low response rate, which is in line with that of other survey years, is a common
phenomenon of voluntary firm surveys in Germany. It reflects the very large number of firm
surveys that target the same firm population. As a result, firms are rather reluctant to
participate in voluntary surveys, causing a low response rate for all of these surveys. Since a
low response rate may cause a bias in the net sample with respect to key variables such as the
share of innovating firms, a comprehensive non-response survey (NRS) was performed. Out
of non-responding firms, a stratified random sample was drawn and firms were contacted by
telephone and questioned on a few key innovation variables (product and process innovations,
R&D activities). The response rate of the NRS was about 85%, and the net size 4,120. Taking
the net sample and the NRS together, the total response rate was 38.4%. While 53.2% of all
firms in the net sample reported to having introduced innovations during 2000 and 2002, this
share was 59.8% in the NRS. When controlling for differences in the size and sector structure
of both samples, there were no statistical differences in the propensity to innovate between the
net sample and the NRS.
We restrict our analysis to firms with less than 250 employees, following the standard
definition of small and medium-sized enterprises (SMEs) as used by the European
Commission. 3,602 firms of the net sample (= 79.4%) are SMEs, of which 1,715 (= 47.6% of
all SMEs) introduced either product or process innovations during 2000 to 2002. Since not all
firms provided full information on all model variables, the number of SMEs available for
model estimations reduces to 2,841, of which 1,049 are innovators. These firms constitute the
empirical base of this study. The number of innovators is reduced stronger than that of non-
innovators since we require more information from innovators, particularly on their
innovation management practice and their innovation success. The size and sector structure of
innovators without full information is not statistically different from the one of innovators
with full information (see table A1).

21
 All variables related to innovation activities, including the dependent variable, R&D input
and innovation management practices refer to a three year reference period (2000-2002),
complying with CIS survey methodology. This standard practice in innovation surveys may
imply an endogeneity problem in case the events measured through the success variable took
place prior to R&D activities or the use of certain innovation management practices. To limit
this problem, we consider market novelties, product line novelties, efficiency innovations and
quality innovations for construction the dependent variable SUC only in case a firm reported
positive economic results with the respective type of innovation in 2002 (i.e. positive sales,
positive cost savings or a growth of sales).
Innovation success is measured as a count variable that sums up the occurrence of
successfully introducing market novelties, product line novelties, efficiency innovations and
quality innovations. We tested alternative indices, for example by adding an extra unit to the
index in case of market novelties or if both product and process innovations were introduced.
We also tested a variant of SUC by merging categories 2 and 3 into one category, i.e.
distinguishing between firms with no “challenging” innovations, with only one, with two to
three, and with all four. Using these alternative measures did not alter any of the main
estimation results. We therefore proceeded with the simplest variant of the index. Table A1 in
the Annex shows the number of observations by combination of innovation type that form our
innovation success index. The additional regression and test results can be found in Tables A2
and A3.
Variables on R&D are directly taken from the corresponding questions in the questionnaire.
Each of the four innovation management variables combines a set of separate questionnaire
items. The survey contained a set of nine items on different human resource management
instruments which are frequently used in businesses to support innovation:6 (1) innovation
output as part of goal agreements with managers, (2) identifying, promoting and committing
individuals who drive innovation processes, (3) recruitment and training of skilled personnel
needed for innovation, (4) delegating decision making of innovation managers, (5) financial
incentives for innovation managers, (6) non-financial incentives for innovation managers, (7)
incentives for employees to develop and report innovation ideas, (8) organisational measures
for a more efficient use of human capital such as innovation circles, (9) engaging employee
representatives in implementing innovations. Firms had to assess the contribution of each

22
instrument to support innovation in their firm on a 3-point Likert scale. Since the items refer
to different firm environments, not all of them are equally relevant to a specific firm, i.e. some
may only be applied in larger firms. We assume that a firm that uses at least one instrument
(including the one stated as free text) with a highly important contribution to in-house
innovation has an effective human resource management of innovation in place. This
procedure seems to be supported by the fact that all nine items on HRM are highly correlated.
Team working is measured through seven items: (1) supporting informal contacts, (2) joint
development of innovation strategies, (3) open communication of innovation ideas, (4) mutual
support for coping with innovation challenges, (5) regular meetings of heads of business units
to discuss innovation-related issues, (6) temporary exchange of personnel in the context of
innovation projects, (7) cross-functional innovation workshops. We construct the TMW
variable in the same way as for HRM, i.e. a firm applying at least one highly important team
work instrument is regarded as having effective cross-functional co-operation in innovation.

Table 5: Model variables: descriptive statistics

Symbol Name Measurement Mean Std. min max
dev.
Variables for all observations with INN = 1 (1,048 obs.)
SUC Innovation Number of different types of innovations (market 1.63 1.10 0 4
Success novelties, product line novelties, efficiency
innovations, quality innovations) that generated
quantitative innovation success in 2002 (index)
RI Permanent Conducting in-house R&D on a permanent basis 0.51 0.50 0 1
R&D 2000-2002 (dummy)
RE External R&D Contracting out R&D 2000-2002 (dummy) 0.35 0.48 0 1
RI*RE Permanent & Conducting in-house R&D on a permanent basis 0.25 0.43 0 1
external R&D 2000-2002 and contracting out of R&D (dummy)
HRM Human At least one out of ten HRM tools was highly 0.55 0.50 0 1
resource important for supporting internal innovation activities
management in 2000-2002 (see text for individual tools) (dummy)
TMW Team work At least one out of eight team working/cross- 0.74 0.44 0 1
functional co-operation tools was highly important
for supporting internal innovation activities in 2000-
2002 (see text for individual instruments) (dummy)
SEA Searching for Any of the following external sources has triggered 0.71 0.45 0 1
external product or process innovations introduced in 2000-
sources 2002: customers, suppliers, competitors, universities
or other public research organisations (dummy)
COP Co-operation Products or processes introduced in 2000-2002 have 0.54 0.50 0 1
agreements been developed in co-operation with external
with external partners, or innovation co-operation agreements with
partners external partners were in place 2000-2002 (dummy)
SIZE Firm size Logarithm of number of employees in 2002 3.50 1.21 0 5.52

6
The authors would like to thank Norbert Janz and Hans-Georg Gemünden for their conceptual contribution in
the design of this question as well as the one on team work and cross-functional co-operation.

23
Symbol Name Measurement Mean Std. min max
dev.
AGE Firm age Logarithm of years since market entry 2.39 0.82 0 5.07
INT Innovation Total expenditure for innovation (including R&D, 0.20 0.93 0 20
intensity acquisition of machinery and external knowledge,
marketing, training, preparatory work for
innovations) in total sales in 2002 (share)
SUB Public Receiving public subsidies for innovation activities in 0.42 0.49 0 1
subsidies 2000-2002 from regional, national or international
governments (dummy)
HUC Human capital Graduates in total number of employees, 2002 (share) 0.33 0.30 0 1
EXP Export ratio Exports in total sales, 2001 (share) 0.15 0.23 0 1
GRP Group Firm is part of an enterprise group (dummy) 0.32 0.47 0 1
EAS East Germany Firm is located in East Germany (dummy) 0.38 0.49 0 1
SEC1 Sector High-tech manufacturing (NACE 24.2, 24.4, 30, 32, 0.13 0.34 0 1
33, 35.3)
SEC2 Sector Medium-tech manufacturing (23, 24.1, 24.3, 24.5, 0.17 0.37 0 1
24.6, 24.7, 29, 31, 34, 35)
SEC3 Sector Manufacturing of intermediaries (20, 21, 26, 27, 28) 0.09 0.29 0 1
SEC4 Sector Manufacturing of consumer goods (15, 16, 17, 18, 0.13 0.34 0 1
19, 22, 25, 36)
SEC5 Sector Banking, insurance, consulting, advertising (65, 66, 0.09 0.29 0 1
67, 74.1, 74.4)
SEC6 Sector Computer-related activities, telecommunication 0.11 0.31 0 1
(64.3, 72)
SEC7 Sector Engineering and R&D services (73, 74.2, 74.3) 0.14 0.34 0 1
SEC8 Sector Others (10-14, 37, 40-52, 60-63, 64.1, 74.5-74.8, 90, 0.14 0.35 0 1
92.1, 92.2)
Variables for all observations with INN = 0 (1,793 obs.)
SIZE Firm size Logarithm of number of employees in 2002 3.06 1.21 0 5.51
AGE Firm age Logarithm of years since market entry 2.58 0.73 0 5.07
HUC Human capital Graduates in total number of employees, 2002 (share) 0.16 0.23 0 1
EXP Export ratio Exports in total sales, 2001 (share) 0.15 0.23 0 1
EAS East Germany Firm is located in East Germany (dummy) 0.40 0.49 0 1
SEC1 Sector High-tech manufacturing (NACE 24.2, 24.4, 30, 32, 0.03 0.17 0 1
33, 35.3)
SEC2 Sector Medium-tech manufacturing (23, 24.1, 24.3, 24.5, 0.07 0.26 0 1
24.6, 24.7, 29, 31, 34, 35)
SEC3 Sector Manufacturing of intermediaries (20, 21, 26, 27, 28) 0.12 0.32 0 1
SEC4 Sector Manufacturing of consumer goods (15, 16, 17, 18, 0.13 0.33 0 1
19, 22, 25, 36)
SEC5 Sector Banking, insurance, consulting, advertising (65, 66, 0.10 0.30 0 1
67, 74.1, 74.4)
SEC6 Sector Computer-related activities, telecommunication 0.03 0.16 0 1
(64.3, 72)
SEC7 Sector Engineering and R&D services (73, 74.2, 74.3) 0.08 0.27 0 1
SEC8 Sector Others (10-14, 37, 40-52, 60-63, 64.1, 74.5-74.8, 90, 0.45 0.50 0 1
92.1, 92.2)

Information on firms’ success in searching for external sources of innovation is taken from
an extensive question on the significance of different innovation sources and its impact on
innovation (see Sofka, 2007; Beise and Rammer, 2006, for more details on this question).
Essentially, firms were asked for five external sources (customers, suppliers, competitors,
universities/other public research organisations, regulation) which have triggered innovation,
i.e. provided innovation impulses that were essential to successfully introduce an innovation.

24
The significance of each source was surveyed separately for product and process innovation.
We construct a dummy variable taking one, if at least one source, out of customers, suppliers,
competitors, universities/other public research organisations was decisive for introducing a
product and process innovation in 2000-2002. We do not consider regulations here since
regulation-led innovation need not coincide with search strategies.
The presence of co-operation agreements in innovation is measured by combining three
questions. Firms with product and process innovation had to indicate whether these
innovations were predominantly developed in-house, in co-operation with external partners or
by others. All firms reporting a predominantly co-operative development of either product
innovations or process innovations are regarded as having an effective co-operation practice.
Since there may still be innovation based on co-operation in firms that predominantly develop
innovations in-house, we also consider firms stating that they were engaged in innovation co-
operation agreements with external partners.
Most control variables were taken directly from corresponding questions in the
questionnaire (the full questionnaire can be found in Rammer et al., 2005). Firm age was
calculated using firm formation data from the Creditreform data base (the largest credit rating
agency in Germany) which also serves as sampling pool for the MIP. Table 5 reports
definitions and descriptive statistics for all model variables. The median firm in the sample
for model estimations has 35 employees and is 12 years old. Mean size (61 employees) and
age (15.4 years) are somewhat higher.
With respect to our key variables, 51% of all firms conduct in-house R&D on a permanent
basis and 35% of all firms contract out R&D. HRM as an innovation management tool is used
effectively by 55%. 74% of firms report that team work and other methods of cross-functional
co-operation are highly important for supporting innovation. 71% of firms have successfully
searched for external sources of innovation, and 54% were engaged in co-operation
agreements or have developed innovations in a co-operative way with external partners.
Correlation coefficients among model variables can be obtained from the authors upon
request.

6 Model Estimation Results

The results for the selection equation on INN* > 0 are similar for all four models estimated.
As expected, the propensity to innovate increased with firm size. This is also the case for the
share of highly skilled employees (HUC). Firms competing internationally, i.e. firms

25
reporting export activity, are also more likely to innovate, which is possibly due to higher
competitive pressure on international markets. Firms located in Eastern Germany innovate
less than Western Germany companies, and younger firms are more like to introduce new
processes or products to the market (see also Huergo and Jaumandreu, 2004). Finally, as one
would expect, there are also significant differences in innovation propensity across industries.
When it comes to correlations across the innovation input and output equations, we only
find weak selection effects. Only in models 2 and 3, the correlation of error terms, RHO, is
significant at the 10% level. That points to the fact that our second stage equation describes
the innovation outcome in a satisfactory way, and that one does not have to worry too much
about an omitted variable bias (which is captured through selection correction). However, as
the selection term is significant in two models, we prefer to report the full models rather than
results of models ignoring potential selectivity.
With regard to the second equation on innovation outcome the baseline model (see table 6,
model 1) shows that both internal R&D and external R&D matter for generating complex
innovations in SMEs. However, we do not find significant interaction effects of internal and
external R&D. This may be due to the fact that we only consider SMEs which may not be
able to realise as significant economics of scope in their R&D as large firms would possibly
do. If R&D projects in SMEs are significantly smaller or less complex than in large firms, and
SMEs have less projects in total, the cross-fertilisation among projects is thus limited.
When we add the innovation management variables (see Model 2), the results concerning
R&D remain robust, and we see an additional positive effect on innovation outcome of human
resource management, successfully searching for external sources of innovation, and co-
operation agreements. Teamwork has no effect, though.
The results are somewhat different when controlling for interaction effects of permanent
R&D and innovation management (Model [3]). For in-house R&D performers, searching for
external sources of innovation and co-operation agreements do matter while for non-R&D
performers, applying each of the four types of innovation management practices increase
innovation success. Effects are stronger for human resource management and team work than
for searching external sources and co-operating in innovation. This result points to different
effects of innovation management in R&D performers and non-R&D performers.

26
Table 6: Ordered Probit models with sample selection
Variable Model 1 Model 2 Model 3 Model 4
Coef. t-value Coef. t-value Coef. t-value Coef. t-value
Innovation success equation
RI 0.335 3.81 *** 0.233 2.64 *** 0.244 1.34 0.302 0.79
RE 0.508 4.44 *** 0.442 3.86 *** 0.445 3.85 *** 0.435 3.70 ***
RI*RE -0.160 -1.10 -0.156 -1.08 -0.156 -1.04 -0.138 -0.91
SIZE (ln) 0.055 1.61 0.052 1.50 0.049 1.42 0.049 1.39
AGE (ln) -0.095 -2.09 ** -0.101 -2.25 ** -0.099 -2.18 ** -0.108 -2.35 **
INT 0.014 0.37 0.008 0.23 0.008 0.21 0.006 0.15
SUB -0.008 -0.10 -0.126 -1.61 -0.136 -1.73 * -0.146 -1.83 *
HRM 0.198 2.83 ***
TMW 0.114 1.40
SEA 0.280 3.73 ***
COP 0.246 3.30 ***
HRM*RI 0.161 1.64 *
TMW*RI -0.047 -0.37
SEA*RI 0.448 3.80 ***
COP*RI 0.269 2.48 **
HRM*(1-RI) 0.222 2.20 **
TMW*(1-RI) 0.229 2.11 **
SEA*(1-RI) 0.173 1.79 *
COP*(1-RI) 0.198 2.00 **
IMP 0001*RI 0.188 0.39
IMP 0010*RI 1.016 2.10 **
IMP 0100*RI -0.063 -0.14
IMP 1000*RI 0.733 0.87
IMP 0011*RI 0.557 1.37
IMP 0110*RI 0.606 1.55
IMP 1100*RI 0.059 0.14
IMP 1001*RI 0.472 0.89
IMP 1010*RI 0.984 1.78 *
IMP 0101*RI 0.034 0.07
IMP 0111*RI 0.728 1.92 *
IMP 1110*RI 0.452 1.20
IMP 1101*RI 0.654 1.60
IMP 1011*RI 0.945 2.20 **
IMP 1111*RI 0.905 2.47 **
IMP 0001*(1-RI) 0.307 1.06
IMP 0010*(1-RI) 0.172 0.83
IMP 0100*(1-RI) 0.468 1.80 *
IMP 1000*(1-RI) 0.681 1.98 **
IMP 0011*(1-RI) 0.786 2.88 ***
IMP 0110*(1-RI) 0.615 2.92 ***
IMP 1100*(1-RI) 0.678 3.10 ***
IMP 1001*(1-RI) 0.121 0.15
IMP 1010*(1-RI) 0.350 1.06
IMP 0101*(1-RI) 0.315 1.10
IMP 0111*(1-RI) 0.738 3.27 ***
IMP 1110*(1-RI) 0.683 3.46 ***
IMP 1101*(1-RI) 0.712 2.93 ***
IMP 1011*(1-RI) 1.091 2.85 ***

27
Variable Model 1 Model 2 Model 3 Model 4
Coef. t-value Coef. t-value Coef. t-value Coef. t-value
IMP 1111*(1-RI) 0.868 4.30 ***
Joint significance of
χ2(7)= 10.20 χ2(7)= 14.07** χ2(7)= 13.04* χ2(7)= 11.87
industry dummies
Selection equation
SIZE (ln) 0.269 10.67 *** 0.268 10.66 *** 0.269 10.67 *** 0.269 10.68 ***
HUC 1.483 11.19 *** 1.483 11.21 *** 1.483 11.21 *** 1.484 11.22 ***
EXP 0.993 6.37 *** 0.995 6.40 *** 0.993 6.39 *** 0.991 6.37 ***
AGE (ln) -0.182 -4.96 *** -0.182 -4.97 *** -0.182 -4.97 *** -0.182 -4.96 ***
EAST -0.143 -2.50 ** -0.139 -2.43 ** -0.140 -2.44 ** -0.139 -2.42 **
GROUP -0.060 -0.96 -0.058 -0.92 -0.058 -0.93 -0.060 -0.96
Intercept -1.589 -11.92 *** -1.591 -11.95 *** -1.591 -11.94 *** -1.592 -11.95 ***
Joint significance of
χ2(7)= 169.76*** χ2(7)= 169.78*** χ2(7)= 169.86*** χ2(7)= 169.90***
industry dummies
RHO 0.166 1.22 0.243 1.83 * 0.233 1.74 * 0.214 1.57
Log Likelihood -2964.92 -2942.29 -2938.90 -2929.86
Notes: All estimations are based on 2,841 observations, of which 1,049 are non-censored observations (INN = 1) used for
success equation. *, **, *** indicate statistical significance at 10%, 5% and 1% respectively. RHO denotes the estimated
error term correlation across the two equations.
Combinations of innovation management practices IMP follow binary coding, with HRM at position one, followed by TMW,
SEA and COP. “0101” thus indicates that a firm applies TMW and COP, but not HRM and SEA.

As certain management practices may not only substitute R&D, but are also substitutes
among each other, we turn to the results of our Model [4]. As a “summary” result from the
combination of significant interaction terms, it turns out that searching for external knowledge
dominates the positive impact of the management practices for R&D performers. For non-
R&D performers, results are more mixed. Strong effects on innovation success are observed
for all practices that combine three or all four innovation management tools as well as for
combining HRM and team work, team work and searching, as well as searching and co-
operating. Non-R&D performers can also yield higher innovation success when applying only
HRM and only team work tools.
In order to investigate the main research question “can non-R&D performers obtain similar
innovation output as R&D performers through a sensible mix of management practices”
further, we conducted Wald tests on whether the coefficient of R&D plus a certain
management practice combination is significantly larger for R&D performers than the output
obtained by non-R&D performers using the same combination of management tools. As can
be seen from table 7, R&D performers achieve a higher index on the dependent variable than
non-R&D performers for the vast majority of management tool combinations, except in two
cases. However, in most cases the difference between the two types of firms is not statistically
significant. We only find that R&D in combination with searching for external innovation
sources outperforms the non-R&D performers significantly, which is in line with the open
innovation paradigm. All three differences that are statistically significant involve searching

28
for external sources. If, however, an R&D performing firm does not use external sources,
there is no combination of management practices that makes an R&D performer better off
than an innovating firm that decided not to conduct own internal R&D.

Table 7: Wald tests on effectiveness of management tools: R&D vs. Non-R&D performers
Variable Test value = β1 + χRn – χNn Chi-squared value (1 df)
HRM 0.18 0.76
TMW -0.03 0.03
SEA 0.52 9.06 ***
COP 0.32 2.46
IMP 0001 0.18 0.19
IMP 0010 1.15 10.07 ***
IMP 0100 -0.23 0.46
IMP 1000 0.35 0.18
IMP 0011 0.07 0.06
IMP 0110 0.29 1.71
IMP 1100 -0.32 1.19
IMP 1001 0.65 0.55
IMP 1010 0.94 3.25 *
IMP 0101 0.02 0.01
IMP 0111 0.29 1.73
IMP 1110 0.07 0.14
IMP 1101 0.24 0.74
IMP 1011 0.16 0.13
IMP 1111 0.34 3.96 **
Notes: *, **, *** indicate statistical significance at 10%, 5% and 1% respectively.
Combinations of innovation management practices follow binary coding, with HRM at position one, followed by TMW, SEA
and COP. “0101” thus indicates that a firm applies TMW and COP, but not HRM and SEA.

With regard to the control variables in the regression models, only few variables affect
innovation success. Young firms are more likely to introduce complex innovations. This
finding adds to Huergo and Jaumandreu’s (2004) result on a higher innovation propensity of
young firms. There are no size effects7, and only weak sector effects in models [2] and [3].
Surprisingly, the share of innovation expenditure in total sales does not affect the innovation
success of innovating SMEs in terms of complex innovations. The same is true for public
subsidies. SMEs having received public money to conduct innovation activities do not show a
higher innovation success (if at all, the variable is negative and marginally significant in
models [3] and [4]). Note that this does not question a general positive effect of R&D
subsidies on innovation in the economy. This is already captured by the covariates on R&D
irrespective of subsidized or not. Our result just shows that the subsidized R&D does not lead
to more complex innovations than privately financed R&D.

29
7 Conclusion
This paper explores the impact of in-house R&D and innovation management on innovation
success in SMEs. Earlier studies have found that in-house R&D, particularly when combined
with the acquisition of external R&D, is a main driver of innovation success. In this study, we
analyse whether SMEs that refrain from in-house R&D can substitute R&D by certain
innovation management practices in order to achieve a similar innovation success, and which
management tools (and their combination) generates the best results. Innovation success is
measured through a categorical variable that captures the extent to which an SME has
successfully introduced “challenging” product and/or process innovations, i.e. innovations
that significantly change the firm’s market position.
Our findings show that continuous R&D activities are a main driver of innovation success in
SMEs, especially when linked to external knowledge sourcing. External sources can be
tapped through different ways, including acquiring external knowledge by contract R&D,
using external innovation sources such as customer, suppliers and universities, or entering
into co-operation agreements with external partners. But firms without in-house R&D
activities can yield a similar innovation success as R&D performers as long as they apply the
right strategy. On the one hand, relying on external R&D seems to be a promising approach,
while occasional R&D - that is to start R&D activities only in case a certain technological
problem has to be solved - is no successful strategy. On the other hand, human resource
management and team work are innovation management tools that can help non-R&D
performing SMEs to gain similar innovation success as R&D performers, especially when
combined with each other or combined with external knowledge sourcing or formal co-
operations with external partners. Combining all four types of innovation management tools
by non-R&D performers is no promising way, however. Focusing on searching external
sources of innovation without in-house R&D is also a less successful strategy.
The results can be viewed from an innovation economics and innovation management point
of view. If SMEs want to generate complex innovations that substantially improve their
competitive position, conducting in-house R&D, i.e. developing technology competence, is
important. For fully exploiting their own technology competence, acquiring external
knowledge through contracting out R&D is particularly helpful. Since SMEs are highly

7
This also holds true when using the log of the number of employees instead of size class dummies, and there
are also no non-linear effects of size.

30
restricted in the scope of developing new knowledge on their own, complementing their own
technology resources with external knowledge widens their opportunities to successfully
transfer R&D results into products and processes. External R&D also allows SMEs to limit
their own risk, better control costs or R&D and specialising on those technology competences
for which they have the best resources. Interestingly, SMEs without in-house R&D are not
likely to catch-up to R&D performers when using external sources for innovation. It seems
that only in-house R&D creates the necessary absorptive capacity to utilize outside
information.
Another main finding is that applying comprehensive innovation management practices
pays off when it comes to other tools than sourcing external knowledge. SMEs that are able to
apply a large set of innovation management tools effectively, including human resource
management, cross-functional team work, and co-operation agreements, yield similar
innovation success as R&D performers. It suggests that comprehensive innovation
management is a type of intangible investment which gives firms a competitive advantage. In
contrast to R&D, spillovers, risk exposure and funding needs are low for this type of
investment and can be handled rather flexibly, making it particularly attractive to SMEs.
Our results have some relevance for innovation policy. First, the strong focus on promoting
in-house R&D often to be found in innovation policy is not fully supported by our study when
it comes to SMEs. First, in-house R&D seems to be particularly effective only if combined
with external knowledge sourcing. Policy initiatives should thus attempt to combine financial
R&D support to SMEs with strengthening the capacities of SMEs to co-operate with other
partners, including links to customers and suppliers. Secondly, innovation policy should also
acknowledge the key role of external R&D in SMEs and also offer financial support to this
type of R&D activity.
Since innovation management can compensate for in-house R&D if applied in the right
way, policy may try to identify likely barriers in SMEs preventing them from effectively
using innovation management practices, particularly human resource management and team
working. Measures to make SMEs familiar with such management tools, for example,
through best practice diffusion, may be another helpful policy approach.

31
8 References
Acs, Z.J., Audretsch, D.B. (1988), Innovation in Large and Small Firms, American
Economic Review 78, 678-690.
Acs, Z.J., Audretsch, D.B. (1991), R&D, Firm Size, and Innovative Activity, in Acs, Z.J.,
Audretsch, D.B. (eds.), Innovation and Technological Change: An International Comparison,
New York: Harvester Wheatsheaf.
Adams, R., Bessant, J., Phelps, R. (2008), Innovation Management Measurement: A
Review, International Journal of Management Reviews 8, 21-47.
Allen, R.C. (1983), Collective Invention, Journal of Economic Behaviour and Organization
4, 1–24.
Amara, N., Landry, R. (2005), Sources of Information as Determinants of Novelty of
Innovation in Canadian Manufacturing Firms: Evidence from the 1999 Statistics Canada
Innovation Survey, Technovation 25, 245-259.
Arora, A., Gambardella, A. (1994), Evaluating technological information and utilizing it
scientific knowledge, technological capability and external linkages in biotechnology, Journal
of Economic Behaviour and Organisation, 24, 91-114.
Bhattacharya, M., and Bloch, H. (2004): Determinants of Innovation, Small Business
Economics, 22(2): 155–162.
Baumol, W.J. (2002), The Free-Market Innovation Machine, Princeton: Princeton
University Press.
Beise, M. (2004), Lead Markets: Country-specific Drivers of the Global Diffusion of
Innovations, Research Policy 33, 997-1018.
Belderbos, R., Carree, M., Diederen, B., Lokshin, B., Veugelers, R. (2004), Heterogeneity
in R&D Cooperation Strategies, International Journal of Industrial Organization 22, 1237-
1263.
Bresnahan, T., Brynjolfsson, E., Hitt, L. (2002), Information Technology, Workplace
Organization and the Demand for Skilled Labor: Firm-Level Evidence, Quarterly Journal of
Economics 117, 339-376.
Brüderl, J., Preisendörfer, P. (1998), Network Support and the Success of Newly Founded
Business, Small Business Economics 10, 213-225.
Cassiman, B., Veugelers, R. (2002), R&D Co-operation and Spillovers: Some Empirical
Evidence from Belgium, American Economic Review 92, 1169-1184.
Cassiman, B., Veugelers, R. (2006), In Search of Complementarity in the Innovation
Strategy: Internal R&D and External Knowledge Acquisition, Management Science 92, 1169-
1184.
Chesbrough, H. (2003), Open Innovation: The New Imperative for Creating and Profiting
from Technology, Cambridge: Harvard Business School Press.
Cohen, W.M. (1995): Empirical Studies of Innovative Activity, in: P. Stonemann (ed.),
Handbook of the Economics of Innovation and Technical Change, Blackwell, Oxford, UK,
182-264.
Cohen, W.M., Levin, R.C., Mowery, D. (1987), Firm size and R&D intensity: a re-
examination, Journal of Industrial Economics 35, 543-563.
Cohen, W.M., Levinthal, D.A. (1989), Innovation and Learning: The Two Faces of R&D,
The Economic Journal 99, 569-596.
Cohen, W.M., Levinthal, D.A. (1990), Absorptive Capacity: A New Perspective on
Learning and Innovation, Administrative Science Quarterly 35, 128-152.
Cooper R.G., Kleinschmidt E.J. (1995), Benchmarking the Firm's Critical Success Factors
in New Product Development, Journal of Product Innovation Management 12, 374-391.

32
 Crépon, B., Duguet, E., Mairesse, J. (1998), Research, Innovation, and Productivity: An
Econometric Analysis at the Firm Level, Economics of Innovation and New Technology 7,
115-158.
Czarnitzki, D. (2006), Research and Development in Small and Medium-Sized Enterprises:
the Role of Financial Constraints and Public Funding, Scottish Journal of Political Economy
53, 335-257.
Czarnitzki, D. und K. Kraft (2008), On the Profitability of Innovative Assets, Applied
Economics, forthcoming.
De Jong, J.P.J., Vermeulen, P.A.M. (2006), Determinants of Product Innovation in Small
Firms: A Comparison Across Industries, International Small Business Journal 24, 587-609.
Freel, M.S. (2000), Do Small Innovating Firms Outperform Non-Innovators?, Small
Business Economics 14, 195-210.
Galbraith, J.K. (1952), American Capitalism. The Concept of Countervailing Power,
Boston: Houghton Mifflin.
Griffin, A., Page, A.L. (1993), An Interim Report on Measuring Product Development
Success and Failure, Journal of Product Innovation Management 10, 291-308.
Griffith, R., Huergo, E., Mairesse, J., Peters, B. (2006), Innovation and Productivity Across
Four European Countries, Oxford Review of Economic Policy 22, 483-498.
Hagedoorn, J. (2002), Inter-firm R&D partnerships - an overview of major trends and
patterns since 1960, Research Policy 31, 477-492.
Hall, B.H. (2002), The Financing of Research and Development, Oxford Review of
Economic Policy 18, 35-51.
Hall, B.H., Griliches, Z., Hausman, J.A. (1986), Patents and R and D: Is there a Lag?,
International Economic Review 27, 265-283.
Hall, B.H., Jaffe, A., Trajtenberg, M. (2001), Market value and patent citations, RAND
Journal of Economics 36, 16-38.
Heckman, J.J. (1979), Sample selection bias as a specification error, Econometrica 47(1),
153-162.
Hidalgo, A., Albors, J. (2008), Innovation Management Techniques and Tools: A Review
from Theory and Practice, R&D Management 38, 113-127.
Högl, M., and Gemünden, H.G. (2001), Teamwork quality and the success of innovative
projects: a theoretical concept and empirical evidence, Organization Science 12, 435-449.
Huergo, E. (2006), The Role of Technological Management as a Source of Innovation:
Evidence from Spanish Manufacturing Firms, Research Policy 35, 1377-1388.
Huergo, E., Jaumandreu, J. (2004), How Does Probability of Innovation Change with Firm
Age?, Small Business Economics 22, 193-207.
Janz, N., Ebling, G., Gottschalk, S., Niggemann, H. (2001), The Mannheim Innovation
Panels (MIP and MIP-S) of the Centre for European Economic Research (ZEW), Schmollers
Jahrbuch - Zeitschrift für Wirtschafts- und Sozialwissenschaften 121, 123-129.
Janz, N., Lööf, H., Peters, B. (2004), Firm Level Innovation and Productivity – Is There a
Common Story Across Countries? Problems and Perspectives in Management 2, 184-204.
Katila, R., Ahuja, G. (2002), Something Old, Something New: A Longitudinal Study of
Search Behavior and New Product Introduction, Academy of Management Journal 45 (6),
1183-1194.
Kim, J., Lee, S.J., Marschke, G. (2004), Relation of Firm Size to R&D Productivity, mimeo,
The State University of New York at Buffalo.
Kleinknecht, A. (1989), Firm Size and Innovation. Observations in Dutch Manufacturing
Industries, Small Business Economics 1, 215-222.
Kleinknecht, A., van Montfort, K., Brouwer, E. (2002), The Non-Trivial Choice between
Innovation Indicators, Economics of Innovation and New Technology 11, 109-121.

33
 Lee, C.-Y., Sung, T. (2005), Schumpeter’s legacy: A New Perspective on the Relationship
Between Firm Size and R&D, Research Policy 34, 914-931.
Levine, D. (1995), Re-inventing the Workplace, Washington: Brookings Institute.
Licht, G., Zoz, K. (2000), Patents and R&D. An Econometric Investigation Using
Applications for German, European and US Patents by German Companies, in Encaoua, D.,
Hall, B.H., Laisney, F., Mairesse, J. (eds.), The Economics and Econometrics of Innovation,
Dordrecht: Kluwer, 307-338.
Lööf, H., Heshmati, A. (2002), Knowledge Capital and Performance Heterogeneity: A Firm
Level Innovation Study, International Journal of Production Economics 76, 61-85.
Love, J.H., Roper, S. (2004), The Organisation of Innovation: Collaboration, Cooperation
and Multifunctional Groups in UK and German Manufacturing, Cambridge Journal of
Economics 28, 379-395.
Love, J.H., Roper, S., Mangiarotti, G. (2006), Organizing Innovation: Complementarities
Between Cross-Functional Teams, DRUID Working Paper No. 06-27, Copenhagen.
Miles, I.D. (2005), Innovation in Services, in Fagerberg, J., Mowery, D., Nelson, R. (eds.),
The Oxford Handbook of Innovation, Oxford: Oxford University Press, 433-458.
Milgrom, P, Roberts, J. (1990), The Economics of Modern Manufacturing: Technology,
Strategy, and Organization, American Economic Review 80, 511-528.
Miranda, A., Rabe-Hesketh, S. (2006), Maximum likelihood estimation of endogenous
switching and sample selection models for binary, ordinal, and count variables, The Stata
Journal 6(3), 285-308.
Mohnen, P., Röller, L.-H. (2005), Complementarities in Innovation Policy, European
Economic Review 49, 1431-1450.
Nijssen, E.J., Frambach, R.T. (2000), Determinants of the Adoption of New Product
Development Tools by Industrial Firms, Industrial Marketing Management 29, 121-131.
Parisi, M. L., Schiantarelli, F., Sembenelli, A. (2006), Productivity, Innovation and R&D:
Micro Evidence for Italy, European Economic Review 50, 2037-2061.
Peters, B. (2008), Innovation and Firm Performance. An Empirical Investigation for
German Firms, ZEW Economic Studies 38, Heidelberg: Physica.
Plehn-Dujowich, J. (2007), Innovation, Firm Size, and RD Search, Economics Bulletin 12,
1-8.
Rammer, C., Peters, B., Schmidt, T., Aschhoff, B., Doherr, T., Niggemann, H. (2005),
Innovationen in Deutschland – Ergebnisse der Innovationserhebung 2003 in der deutschen
Wirtschaft, ZEW Wirtschaftsanalysen 78, Baden-Baden: Nomos.
Ritter, T., Gemünden, H.G. (2004), The Impact of a Company’s Business Strategy on its
Technological Competence, Network Competence and Innovation Success, Journal of
Business Research 57, 548-556.
Rogers, M. (2004), Networks, Firm Size and Innovation, Small Business Economics 22,
141-153.
Roper, S., Love, J.H., Du, J. (2007), The Limits of Open Innovation: Openness and (quasi-
)Markets in the Organization of Innovation, Aston Business School Research Paper 0713,
Birmingham.
Rosenberg, N. (1990), Why do firms do basic research (with their own money)? Research
Policy 19, 165-174.
Santarelli, E., Sterlacchini, A. (1990), Innovation, Fomral vs. Informal R&D, and Firm Size:
Some Evidence from Italian Manufacturing Firms, Small Business Economics 2, 223-228.
Shipton, H., Fay, D., West, M.A., Patterson, M., Birdi, K. (2005), Managing People to
Promote Innovation, Creativity and Innovation Management 14, 118-128.
Siegel, D.S., Wessner, C., Binks, M., Lockett, A. (2003), Policies Promoting Innovation in
Small Firms: Evidence from the U.S. and U.K., Small Business Economics 20, 121-127.

34
 Soete, L. (1979), Firm Size and Inventive Activity. The Evidence Reconsidered, European
Economic Review 12, 319-340.
Song, X.M., Montoya-Weiss, M.M., Schmidt, J.B. (1997), Antecedents and consequences
of cross functional co-operation: a comparison of R&D, manufacturing and marketing
perspectives, Journal of Product Innovation Management 14, 35-47.
Tidd, J., Bessant, J., Pavitt, K. (2005), Managing Innovation. Integrating Technological,
Market and Organizational Change, 3rd edition, Chichester: Wiley.
Tiwari, A.K., Mohnen, P., Palm, F.C., Schim van der Loeff, S. (2007), Financial Constraint
and R&D Investment: Evidence from CIS, UNU-MERIT Working Paper 2007-011,
Maastricht.
Van Dijk, B., Den Hertog, R., Menkveld, B., Thurik, R. (1997), Some New Evidence on the
Determinants of Large- and Small-Firm Innovation, Small Business Economics 9, 335-343.
Von Hippel, E. (1988), The Sources of Innovation, New York: Oxford University Press.
Zeller, C. (2002), Project teams as means of restructuring research the pharmaceutical
industry, Regional Studies 36, 275-289.

35
Annex

Table A1: Dependent variable: Number of observations by type of innovation

Type of innovation Number of observations Inno-
Market Product line Efficiency Quality Total sample1) Model sample2) vation
novelty novelty innovation innovation # % # % index
0 0 0 0 257 17.2 166 15.3 0
1 0 0 0 153 10.4 112 10.6 1
0 1 0 0 155 10.6 111 10.5 1
0 0 1 0 61 3.4 47 3.5 1
0 0 0 1 78 5.9 56 5.9 1
1 1 0 0 255 17.0 196 17.9 2
0 0 1 1 130 9.8 90 9.5 2
1 0 1 0 11 0.6 8 0.7 2
1 0 0 1 39 2.8 27 2.7 2
0 1 1 0 19 1.2 15 1.2 2
0 1 0 1 41 2.9 25 2.5 2
1 1 1 0 20 0.8 16 1.0 3
1 1 0 1 72 5.6 58 6.3 3
1 0 1 1 37 2.7 29 2.9 3
0 1 1 1 32 2.3 20 2.1 3
1 1 1 1 91 6.8 73 7.5 4
Total 1,451 100.0 1,049 100.0
1) Firms with less than 250 employees having introduced product or process innovation in 2000 to 2002.
2) Firms with less than 250 employees having introduced product or process innovation in 2000 to 2002 with full information on all model
variables.

Table A2: Ordered Probit models with sample selection and reduced number of categories (4
instead of 5 categories)
Model A Model B Model C Model D
Variable Coef. t-value Coef. t-value Coef. t-value Coef. t-value
Innovation success equation
RI 0.301 3.33 *** 0.197 2.15 ** 0.131 0.70 0.166 0.43
RE 0.484 4.09 *** 0.416 3.51 *** 0.425 3.55 *** 0.414 3.40 ***
RI*RE -0.121 -0.81 -0.113 -0.75 -0.130 -0.84 -0.111 -0.71
SIZE (ln) 0.043 1.22 0.037 1.02 0.035 0.97 0.034 0.93
AGE (ln) -0.114 -2.43 ** -0.119 -2.57 *** -0.119 -2.55 ** -0.130 -2.74 ***
INT 0.010 0.25 0.003 0.08 0.002 0.05 -0.001 -0.03
SUB -0.008 -0.11 -0.122 -1.51 -0.133 -1.64 * -0.143 -1.73 *
HRM 0.205 2.83 ***
TMW 0.157 1.86 *
SEA 0.255 3.30 ***
COP 0.236 3.08 ***
HRM*RI 0.174 1.71 *
TMW*RI 0.058 0.44
SEA*RI 0.404 3.32 ***
COP*RI 0.287 2.55 **
HRM*(1-RI) 0.223 2.15 **
TMW*(1-RI) 0.234 2.10 **
SEA*(1-RI) 0.157 1.59
COP*(1-RI) 0.175 1.72 *

36
 Model A Model B Model C Model D
Variable Coef. t-value Coef. t-value Coef. t-value Coef. t-value
Innovation success equation
IMP 0001*RI 0.174 0.35
IMP 0010*RI 1.024 2.04 **
IMP 0100*RI -0.026 -0.06
IMP 1000*RI 0.800 0.94
IMP 0011*RI 0.576 1.39
IMP 0110*RI 0.675 1.69 *
IMP 1100*RI 0.238 0.55
IMP 1001*RI 0.551 1.01
IMP 1010*RI 0.889 1.56
IMP 0101*RI 0.191 0.38
IMP 0111*RI 0.834 2.15 **
IMP 1110*RI 0.556 1.45
IMP 1101*RI 0.823 1.96 **
IMP 1011*RI 0.913 2.07 **
IMP 1111*RI 0.999 2.68 ***
IMP 0001*(1-RI) 0.263 0.89
IMP 0010*(1-RI) 0.131 0.63
IMP 0100*(1-RI) 0.435 1.64 *
IMP 1000*(1-RI) 0.653 1.85 *
IMP 0011*(1-RI) 0.742 2.65 ***
IMP 0110*(1-RI) 0.628 2.91 ***
IMP 1100*(1-RI) 0.710 3.18 ***
IMP 1001*(1-RI) 0.189 0.23
IMP 1010*(1-RI) 0.302 0.90
IMP 0101*(1-RI) 0.261 0.89
IMP 0111*(1-RI) 0.709 3.08 ***
IMP 1110*(1-RI) 0.651 3.24 ***
IMP 1101*(1-RI) 0.675 2.70 ***
IMP 1011*(1-RI) 1.175 3.00 ***
IMP 1111*(1-RI) 0.809 3.93 ***
Joint significance of
χ2(7)= 11.15 χ2(7)= 15.59** χ2(7)= 14.95** χ2(7)= 13.31*
industry dummies
Selection equation
SIZE (ln) 0.269 10.67 *** 0.268 10.66 *** 0.269 10.67 *** 0.269 10.67 ***
HUC 1.482 11.18 *** 1.481 11.18 *** 1.482 11.19 *** 1.483 11.20 ***
EXP 0.994 6.37 *** 0.996 6.41 *** 0.995 6.40 *** 0.993 6.38 ***
AGE (ln) -0.183 -4.98 *** -0.183 -4.99 *** -0.183 -4.99 *** -0.183 -4.98 ***
EAST -0.143 -2.50 ** -0.139 -2.43 ** -0.139 -2.43 ** -0.139 -2.41 **
GROUP -0.061 -0.97 -0.059 -0.95 -0.059 -0.95 -0.061 -0.98
Intercept -1.588 -11.91 *** -1.588 -11.92 *** -1.589 -11.93 *** -1.590 -11.94 ***
Joint significance of
χ2(7)= 169.85*** χ2(7)= 169.92*** χ2(7)= 169.96*** χ2(7)= 169.98***
industry dummies
RHO 0.168 1.18 0.247 1.78 * 0.244 1.74 * 0.219 1.52
Log-Likelihood -2693.08 -2671.39 -2669.17 -2659.78
Notes: All estimations are based on 2,841 observations, of which 1,049 are non-censored observations (INN = 1) used for
success equation. *, **, *** indicate statistical significance at 10%, 5% and 1% respectively. RHO denotes the estimated
error term correlation across the two equations.
Combinations of innovation management practices IMP follow binary coding, with HRM at position one, followed by TMW, SEA and COP.
“0101” thus indicates that a firm applies TMW and COP, but not HRM and SEA.

Table A3: Wald tests on effectiveness of management tools: R&D vs. Non-R&D performers
(regressions with reduced number of categories as shown in Table A2.)

37
 Variable Test value = β1 + χRn – χNn Chi-squared value (1 df)
HRM 0.08 0.14
TMW -0.05 0.06
SEA 0.38 4.50 **
COP 0.24 1.38
IMP 0001 0.08 0.03
IMP 0010 1.06 7.87 ***
IMP 0100 -0.30 0.74
IMP 1000 0.31 0.14
IMP 0011 -0.01 0.01
IMP 0110 0.21 1.71
IMP 1100 -0.31 0.84
IMP 1001 0.53 1.05
IMP 1010 0.75 1.96
IMP 0101 0.10 0.05
IMP 0111 0.29 1.61
IMP 1110 0.07 0.14
IMP 1101 0.31 1.14
IMP 1011 -0.10 0.05
IMP 1111 0.36 4.08 **
Notes: *, **, *** indicate statistical significance at 10%, 5% and 1% respectively.
Combinations of innovation management practices follow binary coding, with HRM at position one, followed by TMW, SEA
and COP. “0101” thus indicates that a firm applies TMW and COP, but not HRM and SEA.

38